"Memory and feasibility indicators in GRASP for Multi-Skill Project Scheduling with Partial Preemption"
Oliver Polo-Mejia, Christian Artigues, Pierre Lopez and Lars Mönch

This paper describes a GRASP algorithm aiming to solve a new scheduling problem known as the Multi-Skill Project Scheduling Problem with Partial Preemption, in which not all resources are released during preemption periods. We use a self-adaptive strategy for fixing the cardinality of the restricted candidate list in the greedy phase of the GRASP. We also propose an adaptive evaluation function that includes memory-based intensification, exploiting the characteristics of the best solutions, and a feasibility element for increasing the number of feasible solutions visited. Numerical experiments show the interest of the proposed approach.